Reading Σ in Statistics Formulas

You are staring at a formula that has Σx in it, and you are supposed to compute the mean. The symbol is the Greek capital sigma, and it means one thing: add up the numbers. That is the entire trick. The notation is a compact instruction, nothing more. Summation notation statistics appears in the formulas for mean, variance, and standard deviation, the three places where intro students get stuck. The source for all notation here is OpenStax Introductory Statistics 2e, sections 2.5-2.7.

The mean is the sum of all data values divided by how many there are. When you read x̄ = Σx / n, the Σx part means take every xᵢ in your data set, add them up, and you are done. The index runs from i=1 to i=n, but you never write the index out when the formula is simple. If your data set is {2, 4, 6, 8}, then Σx = 2 + 4 + 6 + 8 = 20. The mean is 20 / 4 = 5.

OpenStax Introductory Statistics 2e uses the index i as the default dummy variable, but j or k works exactly the same way, the variable name does not change the sum. The lower bound is always 1 in a raw data formula, and the upper bound is n, your sample size. Do not invent a different lower bound unless the problem gives you one.

The failure case: you start summing from zero or from a value outside your data set. The index tells you which positions to use. If your data set has five numbers, the index runs 1 through 5, not 0 through 4. Off-by-one errors are the most common mistake here.

This is the single most common confusion in statistics. The notation Σx² means square each x value first, then add the results. The notation (Σx)² means add all x values first, then square the total. They are not the same number, and using one when the formula calls for the other breaks every calculation that follows.

For the data set {2, 4, 6, 8}:
Σx² = 2² + 4² + 6² + 8² = 4 + 16 + 36 + 64 = 120.
(Σx)² = (2 + 4 + 6 + 8)² = 20² = 400.

The difference between Σx² and (Σx)² is not subtle, 120 versus 400 is a factor of more than three. In variance and standard deviation formulas, that difference determines whether your spread measurement is plausible or completely wrong. A student who confuses them will report a standard deviation that is off by a factor that grows with the data set size.

An easy check: write out the terms. If the expression is inside the Σ, the operation happens per term before the addition. If the expression is outside the Σ, the addition happens first and the operation happens on the total. The sum of x squared vs sum of x all squared is a order-of-operations trap that catches everyone once.

The variance formula uses Σ(x - x̄)². This means subtract the mean from each data value, square the difference, then add up all those squares. The order is: subtract, then square, then sum. Doing the subtraction inside the Σ means the operation is applied to each term before the sum.

For the sample variance formula: s² = Σ(x - x̄)² / (n - 1). The numerator is a single number, the sum of squared deviations, and the denominator is the sample size minus one. If you have the population variance formula, the denominator is N instead.

OpenStax Introductory Statistics 2e is explicit: the notation for sum of squared deviations is Σ(x - x̄)². The mean x̄ is computed first from the same data, then each deviation is found. The sum of those deviations, before squaring, is always zero, that is a property of the mean, not a calculation error. Squaring removes the sign and gives you a positive number that measures spread.

The common sigma notation mean formula used in variance: the term (xᵢ - x̄) appears inside the Σ. The index i runs from 1 to n, and xᵢ is the ith data value. The mean x̄ is a constant for the entire sum, but it stays inside the parentheses because it is part of the expression being squared.

Summation in variance formula: you are summing n terms, each one a squared deviation. The number of terms is always the sample size, and the sum is always non-negative. A negative sum of squares means you made an arithmetic mistake, check the sign of each deviation before squaring.

When you move to correlation and regression, you encounter Σxy, the sum of the products of paired values. If you have paired data (x₁, y₁), (x₂, y₂), …, (xₙ, yₙ), then Σxy means multiply each xᵢ by its corresponding yᵢ, then add the results. The index runs from 1 to n, and the summand is xᵢyᵢ.

For the same five-value data set used in a regression problem: Σxy is not Σx times Σy. The sum of products is different from the product of sums, just as Σx² differed from (Σx)². The distinction matters for the covariance formula, which uses Σ(xᵢ - x̄)(yᵢ - ȳ), the sum of products of deviations.

OpenStax Introductory Statistics 2e uses Σxy in the formula for the Pearson correlation coefficient. The calculation is: r = [Σ(xᵢ - x̄)(yᵢ - ȳ)] / √[Σ(xᵢ - x̄)² · Σ(yᵢ - ȳ)²]. Every Σ in that expression follows the same rule: sum of terms, not term applied to a sum.

The failure case: a student tries to shortcut by computing Σx and Σy separately, then multiplying those totals, and putting that into the correlation formula. That gives the wrong number every time because the pairing is lost. Each pair must be multiplied before summing.

Take the data set: {3, 7, 2, 9, 5}. n = 5.

Step 1: Compute Σx and the Mean

Σx = 3 + 7 + 2 + 9 + 5 = 26. x̄ = Σx / n = 26 / 5 = 5.2.

Step 2: Compute Σx² and (Σx)²

Σx² = 3² + 7² + 2² + 9² + 5² = 9 + 49 + 4 + 81 + 25 = 168.
(Σx)² = 26² = 676.

The difference is 168 vs. 676, they are not close. If a formula called for Σx² and you used 676, every following calculation would be wrong.

Step 3: Compute Σ(x - x̄)² for Variance

First find each deviation:
3 - 5.2 = -2.2, square = 4.84
7 - 5.2 = 1.8, square = 3.24
2 - 5.2 = -3.2, square = 10.24
9 - 5.2 = 3.8, square = 14.44
5 - 5.2 = -0.2, square = 0.04

Sum of squared deviations: 4.84 + 3.24 + 10.24 + 14.44 + 0.04 = 32.80. This is Σ(x - x̄)².

Step 4: Compute Sample Variance and Standard Deviation

Sample variance s² = 32.80 / (5 - 1) = 32.80 / 4 = 8.20. Sample standard deviation s = √8.20 ≈ 2.86.

The table below shows the same work in a format you can use to check your own calculations.

Term (i)	xᵢ	xᵢ²	xᵢ - x̄	(xᵢ - x̄)²
1	3	9	-2.2	4.84
2	7	49	1.8	3.24
3	2	4	-3.2	10.24
4	9	81	3.8	14.44
5	5	25	-0.2	0.04
Σ	26	168	0	32.80

OpenStax Introductory Statistics 2e, sections 2.5-2.7, defines Σx, Σx², Σ(x - x̄)², and Σxy with the index i, lower bound 1, and upper bound n. The sample mean formula is x̄ = Σx / n, and the sample variance formula is s² = Σ(x - x̄)² / (n - 1). Those are the exact formulas you will use in every intro statistics course.

The Σ symbol itself was introduced by Leonhard Euler in 1755, in his work Institutiones calculi differentialis. Florian Cajori, in A History of Mathematical Notations, vol. 2, page 265, documented that Euler chose the Greek capital sigma because it is the first letter of the Latin word 'summa,' which was the standard abbreviation for 'sum' in manuscripts. The common story that sigma is the first letter of the Greek word for 'sum' is false, the Greek word is 'athroisma,' not sigma. The index is always a dummy variable; rename it without changing the sum.

You compute Σx² by squaring each x value, add them up, and get a number. Then you compute (Σx)² by adding the x values and squaring the total. They are not the same, and no formula ever replaces one with the other. When you see a variance formula that contains Σx², do not substitute (Σx)². Write out the terms for a small data set first, three numbers are enough to confirm you have the order correct, and then scale up to the full data set. That check takes thirty seconds and prevents the mistake that causes the most homework errors in intro statistics.